Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add cosmix/loom --skill loom-serializationgit clone --depth 1 https://github.com/cosmix/loomWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/cosmix/loom/loom-serialization)<a href="https://agentmods.dev/skills/cosmix/loom/loom-serialization"><img src="https://agentmods.dev/badge/skills/cosmix/loom/loom-serialization.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00013 | $0.04179 |
| Opus 5 | $0.00006 | $0.02090 |
| Sonnet 5 | $0.00003 | $0.00836 |
| Haiku 4.5 | $0.00001 | $0.00418 |
Grade A, and why
loom-serialization scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 221 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Serialization
Overview
Converting in-memory data to bytes for storage/transport and back. The hard problems are not encode/decode calls — they are schema evolution (old and new code exchanging data safely), format choice (self-describing vs schema'd, text vs binary), and correctness traps (int64 precision, NaN, non-deterministic output, unknown-field handling). This skill is mechanism-level; assume you can call the library.
Format Selection
| Format | Schema | Self-describing | Cross-lang | Notes / when to reach for it |
|---|---|---|---|---|
| JSON | none | yes | yes | Debuggable, ubiquitous. No int64/binary/date types; slow; large. Default for public HTTP APIs. |
| Protobuf | IDL (.proto) |
no (tag numbers only) | excellent | Compact, tag-based evolution, gRPC. Needs .proto to read bytes. |
| Avro | required (writer+reader) | schema travels or via registry | good | No per-field tags/names in payload → smallest tagged binary; schema-resolution evolution. Kafka/Hadoop. |
| MessagePack | none | yes | yes | "binary JSON": same data model, ~2x smaller/faster. Dynamic data without a schema. |
| CBOR (RFC 8949) | none | yes | yes | IETF-standard MessagePack cousin; has a canonical form + tags. COSE/WebAuthn/DTLS. |
| FlatBuffers / Cap'n Proto | IDL | no | good | Zero-copy: read fields without parsing; mmap-able. Games, low-latency IPC. |
| bincode / postcard | Rust type layout | no | no (Rust↔Rust) | Fastest+smallest for Rust-only. No evolution — field order/type is the contract. |
| TOML / YAML | none | yes | yes | Config, human-authored. Not for hot-path data (YAML esp. slow + footguns). |
Decision heuristics: public API / must be human-readable → JSON; internal RPC, polyglot, evolving → protobuf (gRPC); event stream with a registry → Avro; Rust-to-Rust cache/IPC, no evolution → bincode/postcard; need to read one field of a huge blob → FlatBuffers; cryptographic canonicalization → CBOR canonical or JCS.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 221 lines · 13 tokens per session scan A c4b7c55ade23
loom-serialization is a skill published in the GitHub repository cosmix/loom (54 stars, last pushed yesterday), licensed MIT. It adds 13 tokens to every session and 4,179 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other skills, from other repositories
API Discoverability for Agents
Making self-hosted services agent-discoverable — bake in a machine-readable API description (OpenAPI spec or a minimal API.md) when building, and discover-first (spec paths, repo search) before probing when integrating.
sandbox-next
Build or maintain Cloudflare Sandbox apps on @cloudflare/sandbox@next (SDK 1.0 preview). Use sandbox-migrate-to-next when porting a stable app.
sandbox-stable
Build or maintain Cloudflare Sandbox apps on the stable @cloudflare/sandbox package. Use sandbox-next for preview apps and sandbox-migrate-to-next for stable-to-preview migrations.
durable-objects
Build, debug, or review Cloudflare Durable Objects code for persistent state and coordination.
cloudflare
Discover and choose Cloudflare products for apps, APIs, AI agents, storage, networking, and security. Use for architecture and product selection, including when the user describes a need without naming a Cloudflare product; then find the relevant skill or documentation.
cloudflare-email-service
Implement or troubleshoot Cloudflare Email Sending and Email Routing integrations and their delivery configuration.